Google has announced that it is expanding its ties with the Pentagon for AI use. The company will now allow personnel to create custom AI agents using Gemini via the Pentagon’s enterprise AI portal, GenAI.mil. The announcement comes at a time when the US Department of Defense is at odds with Anthropic with the latter filing a lawsuit after being designated as a supply chain risk.
In a blog post, Google announced that 3 million civilian and military personnel will have access to custom agents via the portal. Users will be able to design their own agents for tasks such as creating drafts and planning projects. However, it is restricted to unclassified work.
US Defense Undersecretary Emil Michael told Bloomberg, “We’re starting with unclassified because that’s where most of the users are, and then we’ll get to classified and top secret.”
How will this work?
According to Google, users will not need to have prior knowledge of coding. Rather, they can create custom AI agents via natural language prompts through the Agent Designer. The company claims that this will allow users to “design digital assistants to automate repetitive, multi-step administrative tasks”
Google had first provided access to Gemini via the GenAI.mil portal in December 2025.
Pentagon confident Google will be a good partner after Anthropic fiasco
The Pentagon has faced scrutiny recently after it terminated Anthropic’s contract over disagreements on AI use. The Dario Amodei-led firm wanted specific safeguards to prevent mass domestic surveillance or the development of autonomous weapons. On the other hand, the US Defense Department asserted that it wanted access to AI for “all lawful purposes”
The fallout has grown deeper since. While the Pentagon did reach an agreement with OpenAI, it has designated Anthropic as a supply chain risk. With Amodei’s firm has sued the Pentagon.
However, Emil Michael believes that Google will be a reliable partner as the department plans to use its AI agents on classified work in the future. He explained, “I have high confidence they’re going to be a great partner on all networks.”
Pentagon sees no solution to Anthropic fallout
While some reports indicate that Anthropic may be in discussions with the Pentagon to come to a new agreement, Emil Michael believes that the two sides are unlikely to resolve this dispute. He said, “The talks are over. We’re moving on.”
Since Anthropic’s termination, the company’s Claude chatbot has seen a massive rise in downloads. On the other hand, OpenAI, Pentagon’s new AI provider for classified networks, faced massive backlash with users starting a QuitGPT movement online.
